A Novel Signature of Immune- and Inflammation-Related Genes to Predict Prognosis of Hepatocellular Carcinoma

Abstract

Background: Hepatocellular carcinoma (HCC) progression correlates with immune dysfunction and systemic inflammation, necessitating novel immune-inflammatory biomarkers for prognosis.
Methods: This study integrated transcriptomic and clinical data from TCGA-LIHC and ICGC-LIRI-JP cohorts. Immune-inflammatory differentially expressed transcripts (DETs, |log2FC|>2, FDR<0.05) were identified using limma, followed by univariate Cox analysis. An 8-gene prognostic model was constructed via LASSO-Cox regression. Model validation included Kaplan-Meier analysis, time-dependent ROC (timeROC), PCA/t-SNE (Rtsne), and external ICGC validation. Functional enrichment (GSEA/ssGSEA) assessed pathways. Drug sensitivity correlations were analyzed using NCI-60 data. A nomogram integrating genes and clinical features was established (rms package) and calibrated via bootstrap resampling (1000 iterations).

Results: Eleven immune-inflammatory prognostic DEGs (e.g., SLC7A11, MFAP2) were identified. KCNQ3 showed significant OS impact (HR=3.553, P=0.003). The 8-gene model (DNASE1L3, SLC16A3, etc.) stratified high-risk patients with poor OS (TCGA: log-rank P=9.71e-10; ICGC: P=2.278e-04) and high AUCs (1/3/5-year: 0.791/0.727/0.718 in TCGA; 0.730-0.722 in ICGC). Multivariate Cox confirmed risk score as an independent predictor (TCGA: HR=4.210; ICGC: HR=3.189, both P<0.001), correlating with advanced grade (P=2.7e-07) and stage (P=0.00023). Enriched pathways included cell cycle and PI3K-Akt. DNASE1L3-associated genes conferred chemotherapy resistance, while SLC16A3 enhanced ATO sensitivity. The nomogram demonstrated high 1/2/3-year survival prediction accuracy.

Conclusion: Eight immune-inflammatory genes were identified as pivotal prognostic biomarkers for HCC, offering insights into TME dynamics and personalized therapeutic strategies for chemotherapy/immunotherapy.